This class will teach a method that Ljungqvist and Sargent (Recursive Macroeconomic Theory, Third Edition, MIT Press, 2012) call 'dynamic programming squared'. In this method, a dynamic programming problem contains among its state variables the value function for another dynamic programming problem. This method involves some technicalities that, when mastered, empower an economist to attack a variety of challenging problems. The class will teach DP2 by indicating its power for studying Ramsey and Stackelberg problems; problems where enforcement constraints and limited information impair incentive feasible allocations; and settings where reputation can be a more or less imperfect substitute for commitment. The text for the class will be parts of chapters 19, 20, 22, and 23 of Ljungqvist-Sargent.